Abstract
Abstract: :
Purpose: To examine the incidence of change blindness in older observers and its association with accident rates. Methods: Thirty–two observers over the age of 55 (M= 64.6 ± 8.8 years) determined whether two sequentially presented driving scenes were the same or different. In different scenes, a single object in the scene was altered either by the removal of or a change to the object (e.g. stop sign disappeared or green light turned to red). The initial scene was displayed for 10 seconds and the subsequent scene was displayed until the observer made a response or a maximum of 15 seconds. Data were collected for 32 pairs of different scenes and 4 pairs of identical scenes. The influence of the type of difference (change vs disappearance), the relevance of the difference to driving, the centrality of the altered object (central vs marginal interest), and the effect of motivational set were assessed. Accident rates were self–reported on a post–experimental questionnaire. Results: Overall, changes were detected better than disappearances, alterations to central items were detected better than marginal items, and driver–relevant alterations were detected more often than driver–irrelevant alterations. However, a significant 3–way interaction among these variables was found such that item detection was significantly greater when the item, in addition to being driver–relevant, was also central and disappeared. Motivational set did not differentially affect performance. The accuracy of detection was negatively correlated with the number of at–fault driving accidents, but not with the overall number of accidents. Interestingly, the detection of driver–relevant alterations was marginally predictive of the number of at–fault accidents. Conclusions: Change blindness tasks, both those including driver–relevant and driver–irrelevant alterations, appear to be promising means for predicting older observers’ driving difficulties. That overall error rates, as opposed to only driver–relevant errors, are predictive of vehicle accident rates suggests that change blindness performance may reflect a more general attention processing deficit and hence, may be predictive of other types of difficulties encountered by older adults (e.g. falls or pedestrian accidents).
Keywords: aging: visual performance • scene perception • detection